The tap takes a quarter of a second. The terminal beeps, the screen turns green, and the transaction is over as far as anyone standing at the counter is concerned. What actually happened in that quarter second involves at least five separate organizations, two distinct financial processes separated by a day or more, a set of message formats older than most of the people using them, and a chain of contractual obligations that determines who eats the loss if anything goes wrong. Almost none of this is visible, and most of it only becomes relevant when something breaks.
Understanding the plumbing is not merely trivia. It explains why refunds take days when payments take seconds, why a merchant can charge you before shipping, why disputes work the way they do, and why certain transaction types get flagged when others sail through.
The Quarter Second: Authorization
The first process is authorization, and its only job is to answer a yes-or-no question: will the issuer stand behind this transaction?
When the card contacts the terminal, the chip performs a cryptographic exchange producing a one-time value unique to that transaction. This is the core security improvement of chip payments over magnetic stripes — the data captured is useless for a second transaction, which is why counterfeit card fraud collapsed after chip adoption and why fraud migrated online, where no chip is present.
The terminal packages the transaction data and sends it to the acquirer, the merchant’s payment bank. The acquirer identifies the card network from the leading digits and routes the message onward. The network — the switching layer that connects acquirers to issuers — routes it to the issuing bank. The issuer runs its checks: is the card valid, is there available credit or balance, does the transaction pattern look consistent with the cardholder’s history, does it trip any risk rule. It returns an approval or decline, and the response travels back along the same path.
Total elapsed time is typically under two seconds, most of it network latency. What has happened at this point is important to state precisely: no money has moved. The issuer has placed a hold on funds and made a conditional promise to pay. This is why a pending transaction can vanish, why hotels and fuel stations place holds larger than the eventual charge, and why an authorization can expire unused.
The Next Few Days: Clearing and Settlement
The second process is where funds actually move, and it runs on a completely different clock.
At the end of the day the merchant submits a batch of authorized transactions for capture. The acquirer forwards them to the network, which performs clearing — matching each capture against its authorization and calculating what every participant owes every other participant. Settlement follows: actual funds move between the issuer’s and acquirer’s accounts, typically through a central bank system, and the acquirer credits the merchant’s account minus fees.
Those fees are worth understanding because they explain a great deal of merchant behavior. The largest component is interchange, paid by the acquirer to the issuer and set by the network. Interchange varies by card type, transaction type, and merchant category — premium rewards cards carry higher interchange, which is what funds the rewards, and card-not-present transactions carry more than in-person ones because they are riskier. On top sits the network’s own assessment and the acquirer’s markup. The merchant’s all-in cost commonly lands somewhere in the low single digits as a percentage.
This structure is the reason merchants set card minimums, offer cash discounts, and steer toward lower-cost payment methods. It is also why alternative instruments proliferate. Gift cards, prepaid credit, and closed-loop stored value bypass the interchange structure entirely for redemption, which is part of why merchants promote them — and why a secondary economy has grown around them, with liquidation services such as the operator at https://cardsinyong.isweb.co.kr buying unused stored value from consumers who would rather hold general funds. Those instruments run on entirely different rails from card payments, which matters enormously for consumer protection: a closed-loop balance has no chargeback mechanism because there is no issuing bank standing behind it.
Settlement timing explains refunds. A refund is not a reversal of the original transaction; it is a new transaction traveling the same path in the opposite direction, requiring its own clearing and settlement cycle. The original payment already settled and the funds already left the system. The delay is structural, not obstruction.
Disputes, and Why the Rails Matter to You
The dispute system is the layer most consumers eventually touch, and it exists because the network needs merchants and consumers to trust a system in which neither party can verify the other.
A chargeback is initiated by the cardholder through the issuer. The issuer provisionally credits the cardholder and pushes the liability back through the network to the acquirer, which pushes it to the merchant. The merchant can contest with evidence, and the process can escalate through defined stages with deadlines at each. Underneath the procedure sits a liability framework determining who absorbs a fraudulent transaction — a framework that shifted significantly when chip technology arrived, moving liability toward whichever party failed to support the more secure method.
Three practical consequences follow. First, chargeback rights attach to the rails, not to the merchant, which is why paying by card offers protections that paying by transfer or by gift code does not. Second, dispute windows are finite and typically measured from the statement date, which is why noticing problems promptly matters. Third, the protections exist because the card networks built and enforce them, which is precisely why instruments outside those networks — stored value, direct transfers, digital codes — feel similar to use and behave completely differently when something goes wrong.
The rails are a remarkable piece of infrastructure precisely because they are invisible. A system connecting thousands of banks across dozens of countries, settling continuously, with fraud rates measured in fractions of a percent, hidden entirely behind a beep and a green light. But the invisibility has a cost: consumers cannot easily tell which of their payment methods carry the network’s protections and which merely resemble them. Knowing that authorization and settlement are separate, that fees shape merchant behavior, and that disputes are a property of the rails rather than of the transaction is enough to make that distinction reliably — and it is the distinction that matters most on the day something goes wrong.
